Where to start with AI without putting donor data at risk

Start with one task a person reviews. Send the AI tool only the fields that task needs, never the whole donor database. Use an AI vendor under business terms that keep your inputs out of model training. Put a written AI use policy in place that your board can adopt. Keep your donor database of record exactly where it is.

1. One task a person reviews. We start with a single recurring job, such as a drafted acknowledgment or a funder report, and a person on your team reviews every output before it lands anywhere.
2. Only the fields that task needs. A draft gets the few fields it requires. Never the whole database, and never a record you have marked confidential.
3. Business terms that keep your inputs out of training. Drafting runs on Claude under business terms that keep your inputs out of model training, on an account in your organization's name at the vendor's published rate.
4. A written AI use policy. Every engagement includes a policy your board can adopt: scope, approved uses, data rules, the vendor, who reviews, and how to report a concern. Read the policy.
5. Insurance and confidentiality. LUR Growth carries technology errors and omissions insurance through The Hartford, and a certificate of insurance is available to your board on request. Everyone who touches your data signs your confidentiality agreement first.

We build in accounts you own, on tools you already pay for, such as Google Workspace, Microsoft 365, GoHighLevel, Baserow, and n8n. Your database of record stays in place. When we leave, you keep the admin rights, the documentation, and the workflows.

Latoya Robinson has 20-plus years in the nonprofit sector, 15 of them as a nonprofit executive director, and she is still in the seat. She answers to a board and to donors herself, so she builds AI the way an executive director needs it: careful with people's information, reviewed by a person, and explainable to a board.

Capacity Diagnostic, $2,400, 2 to 3 weeks. We find the first task worth handing to AI, check what data it would touch, and give your board a clear, board-ready read of the risk and the first safe step. The fee credits in full toward a build.

Start with the Capacity Diagnostic · See AI consulting for nonprofits · Nonprofit AI governance